osd: delete one of the repeated op->mark_started in ReplicatedBackend::sub_op_modify_impl

[shun-s]

http://tracker.ceph.com/issues/16572

…o in ReplicatedBackend::sub_op_modify_impl

delete one mark_start event as there are two same op->mark_started in ReplicatedBackend::sub_op_modify_impl

Signed-off-by: shun-s song.shun3@zte.com.cn

[shun-s]

when using ceph daemon osd.{id} dump_historic_op, it will show repeated info like bellow:

"description": "osd_repop(client.195731.0:41820 8.137  \ /b1cb0937\/rbd_data.1a77b238e1f29.0000000000001b66\/head v 3558'216067)",      

              "initiated_at": "2016-06-28 06:08:52.980918",    
              "age": 549.619780,   
              "duration": 9.557757,   
              "type_data": [  
               ....   
                    {  
                        "time": "2016-06-28 06:08:52.981251",
                        "event": "started"
                    },
                    {
                        "time": "2016-06-28 06:08:52.981375",
                        "event": "started"
                    }, 

[shun-s]

@tchaikov would you please take a minute to review this pr, thanks

[tchaikov]

@smithfarm i understand

we don't backport cleanup or features to LTS branches unless they are critical or requisite for backporting bug fixes. and in general, we cherry-pick commits from master.

but this change only exists in hammer, and removes the duplicated entries in the output of dump_historic_op.

personally, i am inclined to include this fix. what do you think?

[smithfarm]

@tchaikov OK with me. I created http://tracker.ceph.com/issues/16572 to track. @dachary What do you think?

[smithfarm]

@shun-s: Please amend the commit message so it includes a line:

Fixes: http://tracker.ceph.com/issues/16572

immediately prior to the "Signed-off-by" line. Thanks!

[shun-s]

@smithfarm sorry for seeing this patient review just now, and amend Fixes: http://tracker.ceph.com/issues/16572 is done, please take a look, thanks

[shun-s]

@smithfarm please take a minute to review, much appreciate

[smithfarm]

LGTM but will only be merged after it gets through integration testing, which is ongoing. Please be patient.

[smithfarm]

@tchaikov This PR is in the latest round of hammer-backports integration tests, which passed a rados run (the only failures are a valgrind false positive that has since been fixed by ceph/teuthology#915 and http://tracker.ceph.com/issues/15139 which is an infrastructure issue with two of the tests) - for details, see: http://tracker.ceph.com/issues/15895#note-18

OK to merge?

[tchaikov]

@smithfarm yeah!